Enhanced Subsurface Defect Detection in Structural Steel Using Frequency Modulated Thermal Wave Imaging

Abstract

Structural health monitoring of metallic structures requires reliable non-destructive evaluation techniques capable of detecting subsurface defects with high sensitivity and spatial resolution. This work presents a comparative study of Linear Frequency Modulated Thermal Wave Imaging (LFMTWI) and Digitized Frequency Modulated Thermal Wave Imaging (DFMTWI) for subsurface defect detection in a structural steel specimen containing artificial defects. The specimen was excited using frequency-modulated thermal signals ranging from 0.01 Hz to 0.1 Hz for 100 s, and the transient thermal responses were processed using Time-Domain Phase Analysis (TDP) and Cross-Correlation Coefficient (CCC) after zero-mean normalization. The obtained phasegrams and thermograms demonstrated improved thermal contrast and defect localization for DFMTWI compared with LFMTWI. Experimental results showed that DFMTWI successfully detected deeper defects that were not clearly visible using LFMTWI. Quantitative analysis confirmed the superior performance of DFMTWI, with TDP-based phasegrams yielding SNR values of 40.881 dB, 40.693 dB, 27.263 dB, and 28.774 dB for defects D1-D4, respectively, while CCC-based thermograms achieved 37.266 dB, 32.106 dB, 25.887 dB, and 26.851 dB. Although TDP provided higher SNR values, CCC offered more consistent performance under varying thermal conditions through correlation-based matched filtering. These results demonstrate the effectiveness of DFMTWI for enhanced subsurface defect characterization in metallic structures.